The proper definition of the numerical model of the analysed structure is one of the key task of structural design. It is essential that the description of numerical models differs as little as possible from the actual behaviour of the structure. For this reason, it is extremely important to be able to verify and calibrate numerical models using experimental studies. This article includes a series of experimental tests and numerical analyses for two types of steel bar domes. The structures differ in height. The aim of this study was to calibrate a numerical model of a dome structure by incorporating relevant geometric nonlinearities and fabrication imperfections, in order to match experimental results. The numerical simulations were conducted in Abaqus using geometrically nonlinear analysis with the Riks method, which enables tracking the full equilibrium path, including buckling and snap-through phenomena. Initial simulation results deviated significantly from experimental observations, particularly in terms of the keystone node displacement and the load-displacement relationship. To determine the source of discrepancies, a detailed investigation was carried out, focusing on boundary conditions, fabrication quality, and support compliance. Slight settlement was observed at the supports, which indicated the need to introduce elastic supports into the numerical model. Additionally, geometric imperfections were implemented in the form of the first buckling mode with a realistic amplitude, in accordance with design standards. The incorporation of both elastic support conditions and geometric imperfections led to a substantial improvement in the correlation between numerical and experimental results. The study demonstrates that even small-scale local effects, such as settlement or member imperfections, can significantly influence the global structural response. The obtained static equilibrium paths for the high-rise and low-rise coverings differ significantly. The linear analysis with linear buckling analysis work well for the high-rise dome. For the low-rise dome, a nonlinear analysis was necessary. These structures are subjected to large displacement gradients, and the analysis of the equilibrium path in the range of geometric nonlinearity is essential.